** Readme file for data materials associated with “Asymptotic Behavior of a t Test Robust to Cluster Heterogeneity" **

This folder contains 4 do files: 
1) CSS program to calculate effective number of clusters:

this is the program file that must be run to implement the post-estimation calculation of the effective number of the clusters. The program name is "effclusters_CSS"

2) CSS simulation do file

this program can be run to create data sets used to illustrate results in the papers (figures 1 through 4). this program creates CSS_sim_testsize.dta, CSS_sim.dta, and groupsizes.dta

3) CSS program to create figures 

this program creates figures from the data sets created by the simulations

4) CSS program to calculate effective number of clusters for empirical examples 

this program calculates the effective number of clusters for the two empirical examples (Krueger 1999 and Hersch 1998) in the paper. The original data sets needed to replicate are listed below along with their original sources

webstar.dta - KRUEGER EXAMPLE - DATA FROM MHE WEBSITE http://economics.mit.edu/faculty/angrist/data1/mhe/krueger

CGM_JBES_2011_hersch.DAT - HERSCH EXAMPLE - data provided by Colin Cameron (was used in JBES paper)


Please contact the author of these programs if anything is unclear:
Kevin Schnepel (kevin.schnepel@sydney.edu.au)
